The circadian rhythm: A key variable in aging?

Abstract

The determination of age-related transcriptional changes may contribute to the understanding of health and life expectancy. The broad application of results from age cohorts may have limitations. Altering sample sizes per time point or sex, using a single mouse strain or tissue, a limited number of replicates, or omitting the middle of life can bias the surveys. To achieve higher general validity and to identify less distinctive players, bulk RNA sequencing of a mouse cohort, including seven organs of two strains from both sexes of 5 ages, was performed. Machine learning by bootstrapped variable importance and selection methodology (Boruta) was used to identify common aging features where the circadian rhythms (CiR) transcripts appear as promising age markers in an unsupervised analysis. Pathways of 11 numerically analyzed local network clusters were affected and classified into four major gene expression profiles, whereby CiR and proteostasis candidates were particularly conspicuous with partially opposing changes. In a data-based interaction association network, the CiR-proteostasis axis occupies an exposed central position, highlighting its relevance. The computation of 11,830 individual transcript associations provides potential superordinate contributors, such as hormones, to age-related changes, as in CiR. In hormone-sensitive LNCaP cells, short-term supraphysiologic levels of the sex hormones dihydrotestosterone or estradiol increase the expression of the CiR transcript Bhlhe40 and the associated senescence regulator Cdkn2b (p15). According to these findings, the bilateral dysregulation of CiR appears as a fundamental protagonist of aging, whose transcripts could serve as a biological marker and its restoration as a therapeutic opportunity.

Keywords: Bhlhe40 (Dec1); aging hallmarks; circadian rhythms; interaction network; machine learning; mouse cohort; organ strain sex; sex hormones.

FIGURE 1

Aging alterations in and across organs. (a) The study design and the number of biological replications in black were compared to two existing approaches in dark (Schaum et al., 2020) and light gray (Tabula Muris, 2020). Black/6‐based (B6) or DBA/2 (D2) strains, female (♀) or male (♂), and the number of mice represented as the length of the line from the lowest (2) to the highest (7) distinct group. (b) t‐SNE for time‐independent characteristics. (c, e) Pairwise comparisons depicting the organ‐specific number of DEGs during aging in a line plot referring to 3 (3 m) as solid or 12 months (12 m) as dashed lines compared to later times. Inset: Scheme for pairwise comparisons for 3 m or 12 m. (d) Line plot displays the number of DEGs during common and distinct aging with respective organ exception (e.g., ‐brain: Brain is excluded from the joint analysis). (f) Unsupervised cluster for the 489 nonrejected features of the variable age using PCA across all samples (n = 700) with color‐coded timing. (g) The corresponding loading plot highlights six variables associated with aging. (h) Age‐dependent violin plots are used for all samples for Npas2, Arntl, Nfil3 and Dbp, Per3, and Bhlhe41. DEG, differentially expressed genes; t‐SNE, t‐distributed stochastic neighbor embedding.

FIGURE 2

Analogous gene expression profiles of biological ontologies among collective organ aging. (a) zScore averaged line plot of the four informative module eigengene (MEs), which were additionally subdivided according to hierarchical cluster to reveal potential antiparallel expression behavior within the MEs. Amount of hits as n. (b) Term description of local network ontologies by strength, number of hits per pathway, and false discovery rate (FDR) enrichment. (c) The upper part of StringDB protein‐based network analysis with clusters according to WGCNA and color ontology mapping. The edge‐weighted spring‐embedded layout arrangement is based on the inverse interaction score (confidence from 0.4 to 1) and automatic overlap removal. The size of nodes is based on the iteration average of the numerical‐dependent variable analysis meanImp (Table S3). Colors are classified in figure (b). (d) The opposing expression profiles ME_r and ME_y are visualized, as in figure (a), with organic arrangements and automatic overlap removal. Primarily encompassing the circadian rhythms (CiR)‐proteostasis axis, clarifying the particular deregulation of the CiR feedback loop. (e) zScore averaged line plot of the unique ME₄. Amount of hits as n.

FIGURE 3

The transcript interaction network of aging, created based on the CrossBoruta results. (a) Center, (b) upper left, (c) right, and (d) lower left accumulation. The shape indicates whether the hit is detected numerically, categorically, or both. The arrow directs from a target variable to its correlated feature. The arrow color matches the normHits' iteration average value (the darker the arrow color, the more confident the feature relation). An edge‐weighted spring‐embedded layout creates the arrangement of nodes based on the inverse meanImp iteration average value (inverse arrow length, the closer, the stronger feature associated). Nodes are colored according to closeness centrality. The frame color depicts log₂ asymmetric association strength (AAS), a ratio between the transcript dependence and its impact on the confirmed features (black, no value). Inset: Simplified network of transcripts (A–D). If a target variable has a feature that associates, an arrow points to this feature. The arrow length is inversely proportional to the strength of the association. The black line between A and B symbolizes a value of the simple bivariate correlation. Transcript A would have a high ASS value, and transcript D would have a low ASS value.

FIGURE 4

Steroids affect specific age‐associated targets. LNCaP with (a) 30 nM dihydrotestosterone (DHT) for 24 h (n _CTL&DHT = 6), (b) followed by a 72 h recovery (n _CTL&DHT = 6), (c) or 30 nM estradiol (E2), respectively 300 nM hydrocortisone (Cor) for 24 h (n _CTL&Cor = 6, n _E2 = 5). The relative log₂ changes, normalized to one compared to the tested stable reference gene Hprt1, are plotted for specific genes of interest. For each candidate, the dimethyl sulfoxide (DMSO) control is shown on the left, and the treatment is on the right. Significance is marked with an asterisk (p > 0.05).